A Framework for Damage Detection under Varying Temperature Effects Using Artificial Neural Networks and Time Series Analysis

Abstract

A new damage detection framework with emphasis on damage detection under varying temperature effects is introduced in this paper in the context of Structural Health Monitoring (SHM). One of the most important issues with continuous SHM is the environmental effects (such as temperature, humidity, wind) on the measurement data, which can produce bigger effects in the response of the structures than the damage itself. Without appropriately considering these factors in the damage detection process, the efficiency of this process may be questionable for practical applications. Temperature is considered as one of the most important and influential environmental effect. In this study, an artificial neural network based approach integrated with a sensor clustering time series analysis is employed for damage detection under the temperature effects. This methodology is applied to a finite element footbridge model, where it is demonstrated that the method can successfully determine the existence, location and extent of the damage for different types of load cases, and with different levels of the noise. doi: 10.12783/SHM2015/96